10 ms·
What Happens At Y Combinator
- philwelch 16y agoSome of my favorite (unintentionally?) humorous bits: We have to recruit the largest founders in each batch to act like Tokyo subway pushers and herd all the investors back to their seats. The founders of Cloudkick are sysadmin gods who within the alumni network play the same role for servers that The Wolf in Pulp Fiction played for the back seats of cars.
- ABrandt 16y agoUpvote for bringing the Tarantino reference to my attention; I missed that one somehow. I'd be hard pressed to find a more perfect analogy for the way I think about server issues. If you're not familiar: http://www.youtube.com/watch?v=wWmRTjLRMfU http://www.youtube.com/watch?v=wWmRTjLRMfU
- kn0thing 16y agoI've also found myself leaning on this analogy to describe the ideal non-technical co-founder. When it comes to dealing with any non-technical problem, you'd damn well better be Mr Wolf and just get shit done.
- someone_here 16y agoHow much larger can Y Combinator get? 36 start ups is already a huge number.
- pg 16y agoI don't know. One of our general principles with YC is to apply software techniques to startup funding. So we approach scaling YC the way you'd approach scaling software: we scale up till we hit a bottleneck, then we remove that bottleneck and keep going. Funding 36 startups at a time would have seemed impossible back when we were doing 8. And in fact it would have been impossible to do 36 the way we operated then. So how many will we be able to fund at once, after we've taken advantage of whatever new techniques we discover for scaling? We can't say, because we haven't discovered them yet.
- davidw 16y agoI'm curious whether, and how, you're going to start pulling in alumni to scale it up, and how that will turn out.
- pg 16y agoYes, in several ways. We hired Harj and Alexis, who are alums. We have some of the alumni read the applications, in case we overlook any good ones. And the alumni who come in to give advice about funding are fabulous.
- davidw 16y agoIn order to be self-replicating, you'd have to have alumni with relatively large exits who then participate on more or less equal terms... is that something you foresee, or is it going to just be too messy to add people that way?
- pg 16y agoI haven't thought too much about it. I usually focus on the bottleneck immediately in front of us. But what we do is pretty parallelizable.
- SkyMarshal 16y agoMaybe an eventual next step is creating a YCombinator for YCombinators - a program to help your successful alumni do what you did and become successful incubators/angels/VCs, while keeping them under the overall YC umbrella.
- pg 16y agoI don't think you need multiple YCs though. All our experience so far suggests it's better to have one big YC. Maybe there's some limit to scale, but the optimistic hacker in me would rather prove that by hitting it than assume it's there.
- pg 16y agoThis was an attempt to describe everything we do, so if there's anything I didn't cover, please let me know. (Sorry it's so long. I was surprised myself how long it turned out, but we do a lot.)
- Scott_MacGregor 16y agoIt was long, but not too long, because it was thorough. I didn't realize you were doing so much. In a way the article is kind of like a long sales pitch for YC without all of the typical sales language hype. Which is nice.
- robg 16y agoBesides music labels, are there other industries you try to steer startups away from, either through not accepting them or through nudging them in other directions? I'm looking specifically at medical applications, but I imagine there are others too.
- pg 16y agoNo others are quite as bad. We'd be dubious about startups building stuff for the government, for example, because there is such a horrible mismatch between startup culture and government culture. But we'd still fund a startup building stuff for the government if they weren't naive about what selling to the government would entail.
- robg 16y agoThanks and agreed. I'm surprised I haven't seen many (any?) medical applications from YC companies given the margins available over consumer applications. Do you think it is the uncertainty of the FDA regulations or just a space that Silicon Valley doesn't tend to mine as frequently? Make something patients want is particularly alluring to me.
- pg 16y agoI would guess the reason is that such applications are foreign to most potential applicants.
- r00k 16y agoTypo: "We make a point of introducing speakers to those who particularly want to them."
- pg 16y agoThanks; fixed.
- sperry 16y agoVery impressive. I hope to someday apply, but being a very novice technical founder, that day may be much farther down the road.
- bkudria 16y agoThis is very interesting, but the topics covered make it seem like YC is only focused on advising companies how to raise more money.
- pg 16y agoHrm, that's not good. Most of the convos are actually about the product. We only switch to talking about investors towards the end of the 3 months, and even then we talk more about the product. Maybe it seems that way because this talks about all the stuff that's common to all the startups, and the things to do with the product are usually more startup-specific than those to do with investors.
- bkudria 16y agoRight - my point isn't: YC only advises on investment topics, but: This overview makes it look like YC only advises on investment topics. I'm the last person to give you writing advice, but maybe some examples would help.
- jon_dahl 16y agoMost of the dinners and office hours are focused on building a great startup, not raising money. There is structure around raising money (angel day, demo day, sequoia, etc.), but it's definitely a minority of the focus. I'd also say that even the fundraising stuff isn't just about raising money. Having a smart angel investor give you feedback on your startup isn't just about raising money; it's also about building a product, a team, and a company.
- bkudria 16y agoAgreed. And, pg has already written plenty about advising startups to build things people want - I just want more :)
- robg 16y agoThat wasn't my takeaway. The emphasis to me was on how the creative process is nurtured and accelerated. I'm actually surprised by how many times the word "event" occurs, but stepping back that makes sense. Events are opportunities for action.
- eduardoflores 16y agoAwesome, I envy you
- smountcastle 16y ago"At the end of Demo Day what the startups usually have is a bunch of potential leads. Our default advice is to do a breadth first search, weighted by expected value. ... There are a few investors for whom it should be zero, and we can tell startups about those too." Interesting. If you know beforehand that those investors' expected value is 0, why invite them?
- pg 16y agoIt would be too much of a deliberate affront to refuse to let them even attend. And not all the leads startups have are from the audience at DDay. Since there are so many investors here, they often get introed to investors before DDay through some other connection. Also, zero is probably an overstatement. It would have been better to say epsilon. Because most of these investors I would take money from if the alternative would be to shut down the company.
- aspir 16y agoHow often do those situations happen (ie- take money from a low quality individual or close the doors)? Startup sucess rattes are very low as a whole, so it seems as if there may be a culture of "vulture" investors in SV to prey on entrepreneurs who just need cash. It'd be good to know the warning signs.
- ekanes 16y agoA gentle request; I think it'd be better to call it DemoDay or something other than abbreviating it to DDay. (I know you mean no disrespect, it's an easy shorthand...)
- fatjonny 16y agoI enjoyed reading it. It is nice to read about what the startups in Y Combinator go through. Minor typos I noticed (since you seem like you want the feedback): missing 'the': "to get the company into best shape possible" spelling of predict: "know the investors but can preditct their reaction to each specific startup" see instead of seem: "no one except the current batch of startups gets to seem them"
- pg 16y agoThanks; fixed.
- jbgregg 16y agoThank you Paul. It is great to have this detailed information that fills in all the gaps I have wondered about.
- YuriNiyazov 16y agoI found the following sentence to be hard to parse: "The general answer is to pick something where the product of how fast it can be built and how excited users would be about it is high."
- Swizec 16y agox = speed-of-implementation(idea) y = usefulness(idea) rank(x,y) = x*y Something like that I think.
- patio11 16y agoThat is accurate parse as written. I think the even more dramatic conclusion suggested by Lean Startup is it is probably max(y1..yN)*C^(90/x), where y1 is the utility of your first idea and x is average number of days it takes you to learn one important thing (a full cycle through the idea, implement, measure, conclude loop). Scary stuff starts happening in any business when you take the limit as x approaches zero. (This came up in client work today.)
- budu3 16y agoWhat is C?
- patio11 16y agoA constant, for simplicity. Call it your lower threshold of the impact one good idea makes on your business. (I have a mental target at C = 1.05 but at YC stages there is almost certainly a company with C = staggering.)
- TheSOB88 16y agoCan you put that in the English? I am too tired to make sense of that math. I hope you do not think I am stupid for not immediately seeing the point of many seembols.
- 16y ago
- andrewacove 16y agoHow frequently do companies grow (hire, add a cofounder) during the YC term? Does that ever happen? Is advising on that ever part of the YC role/conversation?
- pg 16y agoMaybe 15-20% do. And yes, we care a lot about that, and often help not only in advising but in recruiting.
- StavrosK 16y agoDo you generally advise single-founder companies to add cofounders (it does have its benefits), or do you think single founders are fine if they pass the filter? I guess my question is, do you see single-founder companies as missing something (a cofounder), or does it depend on the founder?
- pg 16y agoYes, usually, though we don't insist on it. I think most single founders would be better off with cofounders, but it's not worth taking someone mediocre just to have a cofounder.
- budu3 16y agoHave you ever had any successful single founder company came out of the other end of the YC process?
- danilocampos 16y agoMakes me wonder about how selection criteria shifts based on maturity of the company. What signaling does YC look for in a company whose idea was birthed a week before applying, versus one who applies while their beta is in progress? Is it as simple as "these guys are smart" in both cases? I've met lots of guys way smarter than me with less motivation, so there's got to be more to it.
- pg 16y agoMostly we care about the people, regardless of how far along the company is. Which can be a problem, actually, because startups that have been running for a while generally overestimate the percentage of the work that's behind them. For us the most important thing about the company having been running for a while is that it gives us more information about the founders. E.g. we can ask them "what have you learned from your users?"
- danilocampos 16y agoThough it occurs to me now this may be just the thing PG and friends can't share with us. It's to their benefit to make expectations for a great application very clear. On the other hand, I'm sure there are a handful of red flags and heuristics that they emphatically won't share, simply because they're effective filters. Their disclosure could let people better obscure true problems or falsely inflate perceived value. Reviewing applications after that would be a lot less fun.
- zzleeper 16y agoThanks for the insight Paul. As an outsider, it's quite interesting to understand how (successful) startups get formed. It's funny that in my field (economics/finance) we treat startups as black boxes, and we have no clue about how they work. We argue that they are "crucial to the economy and to growth", but yet we haven't made any progress at understanding [1] them. Do you guys think that we (academy) have any chance at finally understanding/modelling what you do? [1] "Understanding" = Either writing a mathematical model, or test some hypothesis with a few regressions.
- pg 16y agoThere are certainly a lot of important and so far mostly unstudied phenomena there. How much of it you could talk about in conventionally publishable economics papers is another question, though.
- johnglasgow 16y agoPaul was interviewed at CrunchUp this summer, and he briefly talked about what they have learned works and doesn't work. The video can be found on TechCrunch. - 2 founders is best - founders should have known each other for a while before starting the company - team is much more important then idea
- Johngibb 16y agoStays at YCombinator?
- cperciva 16y agoIt's a bit misleading to call these events "dinners" though, because they last half a day. People start to show up for dinners around 6 pm. This seems wrong, unless the YC dinners conclude at 6 AM.
- pg 16y agoI meant half a waking day, cperciva. When someone says "I spent half a day trying to find that bug," they don't mean 12 hours.
- StavrosK 16y agoSometimes we do :(
- cperciva 16y agoI meant half a waking day, cperciva. Huh. That never occurred to me as an interpretation. I wonder if this is a British vs. American vs. Canadian English issue, or if I'm just slower than usual today.
- ghshephard 16y agoI'm from British Columbia originally, Half a Day (in the dinner context) is 12 Hours to me. I'll do a sample of the Indians, Canadians, and Americans in the office and get back to you. This may be another of those "organ-eye-zation" versus "organ-eh-zation" things that pop up form time to time.
- lincolnq 16y agoAmerican here. To claim that something takes me half a day, I would usually interpret that as half a working day, so 4-5 hours. I brew beer, and I tell people that "beer brewing takes a full day" -- if I start brewing at 8:30 then I get done before 5. Yeah I could do some other stuff afterwards. Oh, and it's "organ-eh-zation" through and through. :)
- stevefarnworth 16y ago
- jbhelms 16y agoWhat do you recommend to the founders that can't move to the bay area? Like people who own homes, have a family, etc? I know the YC requires it. I once emailed and asked about commuting because my co-founder and I both live in Sacramento, but the person who answered my email basically said that commuting is not feasible.
- pg 16y agoWe funded some guys from the Central Valley once, with houses and families. They rented a place in SV for 3 months, and they'd go home on weekends to visit their families.
- raminf 16y agoI've always thought of YC as more of an incubator for young 20-somethings trying to get an idea off the ground. The impression has been fortified by the reported small amounts of 'seed' that is given out. I know a lot of people in older age brackets (30s-40s-50s) with a family and Silicon Valley mortgage who have great ideas and experience but wouldn't even consider approaching YC. My one suggestion is if you want to expand your age/experience bracket and get these people to apply, you may want to address issues directly relevant to them. One suggestion is to relabel the investment to something more like a 'stipend' and explain clearly that it's not expected to cover salary and life expenses, but only to help defer company expenses. That way, people know that they need to set aside enough savings to cover their family. The other is to spell out clearly the policy of working on outside projects (to bring in some money) during the YC 3-month period. Leaving it ambiguous keeps out people who need a part-time gig to bring in some money but could spend the rest of their waking time working on a cool product. I know people who work as consultants and make $10K-$25K in a month. To these people, the YC 'investment' is more of a deterrent. The main attraction would be to get wired into the network and tap into the experience YC and its alumni can offer. It would be good if you could clarify whether it makes sense for them to apply or not. Overall though, a good write-up with a decent level of detail. Thanks for putting it up.
- noahth 16y agoIf this hypothetical person doesn't realize that he or she might need some savings to cover family expenses while devoting all or most of their time to a project that isn't generating revenue yet, I am skeptical of that person's ability to successfully start or run a business. Just me? Additionally if someone is wary of losing their great consulting/contracting fees to start the startup, maybe starting a startup is not for them. Or maybe they just need to plan to minimize the risk that they will be taking. Either way, this is not YC's problem to fix, imo.
- nl 16y agoYou may find "Doing Y Combinator in your 30s" useful & interesting: http://zencoder.com/encoder-blog/2010/09/21/doing-y-combinator-in-your-30s/ http://zencoder.com/encoder-blog/2010/09/21/doing-y-combinat...
- 16y ago
- swombat 16y agoIf you were to start YC again, from scratch - i.e. if your brain was teleported into your body 5 years ago - what would you do differently? Would you start it directly in SV instead of Boston? Would you focus more on building the alumni network from day one? There must have been many mistakes along the way, while this article talks mostly about the successes. I'm curious what mistakes you made and avoided... maybe that's a topic for another article though.
- pg 16y agoVery interesting question. I would do it more emphatically, for one. The first couple batches we were pretty tentative, because we didn't know it was going to work. It's painful to consider whether it would have been best to do YC in SV year round from the start. I loved alternating between the two places. I still miss living in Cambridge. And yet the founders in the Boston batches probably were getting short shrift. The startup community in Boston is so much smaller. We could never have organized something like Angel Day; there just aren't enough good angels. We wouldn't even have had enough alumni in Boston, because most of the successful ones used to move to the Valley. You can figure out what we'd do differently by what we've changed. We give startups more money on average, but we don't give them any more for having more than 3 founders. Jessica now strongly encourages vesting, not to protect us, but to give startups a predefined way to deal with founders leaving. Interviews and Demo Day presentations are now much shorter, and I'd keep both short even if it weren't necessary for scale. We look more at the people and less at the idea when making funding decisions. We've added a bunch of protections against good groups falling through the cracks in applications. But we haven't changed that much structurally. We've made a lot of mistakes, but they've mostly been one-offs, like funding startups we shouldn't have, or not funding startups we should.
- jacquesm 16y ago> like funding startups we shouldn't have, or not funding startups we should. Examples?
- guelo 16y agoThe main insight I took from this was the quip about "investing in YC-funded startups is safe in the same way buying name-brand products is". Like many people I sometimes feel ripped off when I pay 50% more for that box of Kellogs cereal or whatever. I imagine the same feeling might be the motivation behind AngelGate.
- shykes 16y agoThere's a difference: YC's brand is not based purely on marketing. It's based on the success rate of past YC companies, which is objectively higher than average. From an angel's point of view, YC is a deal refinery. Raw deals come in. High-quality, refined deals come out. Many investors will gladly pay a premium for those, because much of the early risk has been taken out: do the founders get along? Do they crack under pressure? Can they ship something? Can they adapt to changing circonstances? YC figured out a repeatable, streamlined formula for a process which was previously a black art. They industrialized the gut feeling. And they deserve every bit of recognition they got for it.
- ghshephard 16y agore: "YC figured out a repeatable, streamlined formula for a process which was previously a black art. They industrialized the gut feeling." That's not my understanding. I don't think they have any formula other than "We recognize great entrepreneurs when we see them" - It's not something they can write down in a book and sell to other people - so, in that sense, not industrialized. Still a gut feeling. I could be wrong though - happy to hear otherwise.
- shykes 16y agoYou're leaving out an important part of the equation: the sheer number of deals that went through YC, and the data associated with them. When you know the complete story on 500+ founders, and look at which ones were successful, patterns emerge. These patterns can be used to make better decisions and give better advice. Anybody who's talked to PG will tell you he is a walking database of startup patterns. "I've seen X and Y do this and they failed. Don't do it". I'm sure as YC scales, that database will be externalized to a medium other than PG's brain :) So, yes, there is a formula: it's called data.
- andjones 16y agoHow exportable is YCombinator? Much of YCombinator depends on people being in the same geography. I understand that to be the reason why founders are required to move to Mountain View. Do you think it would be possible to start YCombinator prime elsewhere in the country? If we put together many of the same factors, angels, infrastructure, good mentors, could this be replicated, albeit on a smaller scale?
- adbachman 16y agoHe's written about this in the past. "Can You Buy a Silicon Valley? Maybe." http://paulgraham.com/maybe.html http://paulgraham.com/maybe.html and "How to Be Silicon Valley" http://paulgraham.com/siliconvalley.html http://paulgraham.com/siliconvalley.html. He trends towards, "go to them", though: http://paulgraham.com/startuphubs.html http://paulgraham.com/startuphubs.html It comes down to people, money, and time. Exporting Silicon Valley might be the wrong approach. Another place must have its own culture and value system in order to survive and those qualities can't be borrowed. Boulder, Boston, Austin, and to some extent NYC and a few other east coast (of the US) cities (my heart belongs to Baltimore) have burgeoning scenes, but each is different than SV. Where they don't try to simply mimic, I think they're more likely to find success.
- anamax 16y ago> If we put together many of the same factors, angels, infrastructure, good mentors, could this be replicated, albeit on a smaller scale? You need "enough" of all of the critical factors and then you have to overcome whatever penalty you incur from missing some of the "nice to haves". Considering the odds against success even with all of the stars aligned .... Seriously - unless your customers care where you are, why are you so resistant to going to where the odds of success are highest? It's okay to value things other than success, but you're competing with people who are focussed on success. How are you going to overcome that advantage?
- pg 16y agoThe hardest part to replicate elsewhere would be the angels. You really need them, more than you need VCs, and you can't just call them into being.
- dzlobin 16y agoIf the startup either hasn't decided what to work on Approximately what percentage of startups do you accept that apply with no particular idea?
- pg 16y agoNearly all have some idea or ideas. Some make it clear that they're not wedded to their idea, which is fine with us, actually.
- HectorRamos 16y agoI've been eyeing YC for a while and this post has convinced me to apply. I quit my job three months ago and have been doing consulting work in order to generate enough runway to finally dedicate ourselves fully to our own product. I think we will be done with all of our current gigs by november, so the Winter YC cycle looks VERY attractive right now. We have a couple of ideas, and it seems like we might have enough time to get started on them before YC, but still be flexible enough when the cycle starts.
- nitrogen 16y agoI hope you are able to finish your consulting when you expect, but it's my experience that it almost always takes longer.
- abbe 16y agostays at Y Combinator!!
- jeremydavid 16y agoIf a YC startup was to never be acquired, what sort of dividend structure would you think is ideal? What sort of returns would you consider a "successful" investment?
- hyad 16y agoGreat article! Just reading this finally motivated me enough to create an account here and seriously think about applying for winter 2011. After some reconsideration though, I'm not quite sure I'd be willing to quit my day job just yet. I have a few ideas for startups and one project I've been working on in my spare time that is interesting but nowhere near complete. Would it be worth to apply anyway? Perhaps I could list all the ideas I have or maybe just the best one? I was also thinking about applying to the YCommonApp. But I'd prefer to target my resume to specific companies just because that has worked for me the best in the past. Anyone have advice on what might be a good route to take? YCombinator funding? YCommonApp? Both or neither?
- hyad 16y agoTo reply to my own comment: I probably should have signed up to attend this - www.startupschool.org The deadline was 3 days ago. D'oh! If anyone who had signed up is reading this and has to cancel, let me know if I can take your spot. :)
- gosuri 16y agoStays at Y Combinator
- tgriesser 16y agoIn your essay about the application process you discuss figuring out what the combination of founders are (e.g. two hackers, one business one hacker, two hackers one business...different stages of their careers) and you have seen about 20 or 30 different founder combinations. Have you noticed any particular combination that seems to work better than the others across the board, or does it depend on the startup?
- adlep 16y agoThis may be too direct question to ask, but I will anyway to have some points of reference... 1. What do you - at YCombinator consider to be your best example of a successful start-up? If you can name that company, that would be awesome and appreciated. 2. On the other hand, what do you consider to be a flat out failure? An example of an idea that totally did not work? If you could name this venture, it would be appreciated too.
- pg 16y agoI can't pick favorites, but it's pretty easy to see what the most prominent YC-funded startups are. One good example of a bad idea that I can quote without hurting anyone's feelings is Artix: http://paulgraham.com/bronze.html http://paulgraham.com/bronze.html
- rafaelc 16y agoAre there any downsides to being in Y Combinator?
- pg 16y agoThere are two I can think of: you have to move, and you have to give us some stock. But I'm pretty confident that in both cases what you get in return more than compensates.
- edanm 16y agoGreat article, thanks a lot for posting this. I do have one question: is the goal of all the companies to raise addition funding rounds? Or, do some startups still want to work without raising the extra capital, but rather grow organically (37 signals style). I guess another way to ask is: do you force (or even strongly expect) companies to get extra funding on demo day, do you let them choose (and they always choose to get more funding), or do some startups actually not get extra funding? I ask because, from your article, it seems like most of the latter part of the cycle is geared towards raising capital, and I had always assumed that not raising capital was also a viable option for startups.
- pg 16y agoNo, not all of them. Some are already profitable when they apply, and a few more make it to profitability just on what we give them. But the distinction is not a sharp one: most of the profitable companies would take money on sufficiently good terms. Though I don't mention this explicitly in the article, Demo Day isn't just for raising money. It's a good way to get mindshare too, because the people in the audience are so influential. So even the startups for whom raising money is not a top priority usually try hard to be impressive on DDay.
- pclark 16y ago> Dinner itself happens around 7:15. Everyone eats together at long white tables designed by our architect Kate Courteau. The general atmos is like a modernist version of an Oxford college dining hall, but without a high table seriously?
- barry-cotter 16y ago[1] There was some debate whether to keep this analogy, lest it seem a pretentious comparison, but I kept it after Harj, who went to college there, insisted the atmos was in fact very similar.
- nileshtrivedi 16y agoIf we broadcast or even recorded the talks, the speakers would clam up. How about anonymizing it by removing the names of the startups/founders before publishing ?
- barry-cotter 16y agoIt took less than an hour for someone to figure out the Partner and Venture Firm Steve Blank was referring to in his post that was approximately called "You're Just the Founder". Now admittedly he tagged the post with the name of the company so it was easier, but still. Anonymising information while leaving it usable is hard to the point of (near?)impossibility
- eru 16y agoHow about recording (with permission), and asking in a few years time whether you can publish it? Some issues just need to cool down.
- fredliu 16y agoGreat article! (it surprised me that I read the whole article in one shot without evening noticing it was that long!) One question (maybe a bit sensitive): all the glory of successful YC startups aside, what are the "failure rate" for YC funded startups? What situation would you consider a "total fail", if that ever happened? And another related question is based on which criterion would you consider a YC graduate successful? user base? the amount of money raise? financially break even? profitability? thanks.
- paramendra 16y agoThis is an amazing piece of writing by Graham. If he is as good a coder as he is a writer, he must be really good. http://goo.gl/fb/sfxTi http://goo.gl/fb/sfxTi